“A/B Video Testing” and “Multivariate Testing” are methods of comparing various versions of a video (e.g. an online advertisement), to statistically analyze viewer input about which version of the video is most favored by a viewer (e.g. a consumer).
Although there are A/B testing framework systems for creating and testing changes to the text within a website, there is currently no system to allow a user to easily change and test other variables within a video, such as background music, deleting or switching scenes, etc. For example, any change to a video requires generating an entirely new full-length video with the change, and uploading it to the webpage. It is thus very time consuming and expensive to generate and upload a video for each change that is required to be viewer tested.
There is therefore is a need in online video for a networked computer system and method to automatedly, and randomly, generate a statistical sample of all possible full-length video versions, and electronically transmit them for playback. And then for the automated collection and analysis of viewer data to determine which video version is most favored by a plurality of viewers. The automated collection and analysis of viewer data can also be used to determine a different favored video version for each specific segment (e.g. demographic) of viewers. The networked computer system requires a novel video editor and a novel online video provider comprising at least one server able to randomly create, transcode and electronically transmit versions of full-length videos from a data structure of video elements (e.g. video scenes, cards, overlay, background music, etc.).